Assessment of Macrolide Transport Using PAMPA, Caco

CROATICA CHEMICA ACTA
CCACAA, ISSN 0011-1643, e-ISSN 1334-417X
Croat. Chem. Acta 83 (3) (2010) 323–331.
CCA-3423
Original Scientific Article
Assessment of Macrolide Transport Using PAMPA, Caco-2
and MDCKII-hMDR1 Assays
Danijela Nožinić,a,* Astrid Milić,a Lara Mikac,b Jovica Ralić,a
Jasna Padovan,a and Roberto Antolovićc
a
GlaxoSmithKline Research Centre Zagreb, Ltd., Prilaz baruna Filipovića 29, 10000 Zagreb, Croatia
b
Division of Materials Physics, Ruđer Bošković Institute, Bijenička 54, 10000 Zagreb, Croatia
c
Department of Biotechnology, University of Rijeka, S. Krautzeka bb, 51000 Rijeka, Croatia
RECEIVED JANUARY 19, 2010; REVISED MARCH 21, 2010; ACCEPTED MARCH 22, 2010
Abstract. The transport of commercially available macrolide antibiotics erythromycin, clarithromycin,
roxithromycin, azithromycin, and telithromycin was studied by PAMPA, a model of passive diffusion,
and by two models of active transport, Caco-2 and MDCKII-hMDR1. An involvement of efflux pump
P-glycoprotein (P-gp) in macrolide transport was also examined. Generally, in Caco-2 and MDCKII-hMDR1
assays without P-gp inhibition, a significantly lower apparent permeability (Papp) in apical-to-basolateral
(AB) than basolateral-to-apical (BA) direction was observed [efflux ratio, Pappratio(BA/AB) > 4.2]. Upon
P-gp inhibition, Papp(AB) was increased and Papp(BA) was decreased. The observed Papp values, their
efflux ratios, as well as a significant decrease in efflux ratios caused by P-gp inhibition (> 50 %) qualify
all tested macrolides as P-gp substrates. Results obtained with PAMPA suggest that this assay is not a method of choice for screening of macrolide permeability due to the lack of P-gp efflux pump and/or other
transporters potentially involved in macrolide transport.
Keywords: macrolides, permeability, P-glycoprotein, PAMPA, Caco-2, MDCK
INTRODUCTION
The intestinal epithelium permeability is recognized as a
critical feature that determines the rate and extent of
human absorption and at last, the bioavailability of
drugs. Commonly used in vitro methods to assess intestinal permeability in early drug discovery include: parallel artificial membrane permeability assay (PAMPA),
Caco-2 and Madin-Darby canine kidney (MDCK) cell
assays.1–3 PAMPA represents a rapid, low-cost alternative to cellular models for early ADME screening of
new chemical entities, where a mammalian membrane is
mimicked by various filter-supported lipid membrane
systems.4–7 However, due to the lack of facilitators of
active transport, only passive diffusion is assessed by
PAMPA. Alternatively, cellular models offer several
advantages over PAMPA due to over expression of the
transporter of interest, precise control of experimental
conditions, and, usually, the lack of other confounding
physiological processes. Originally derived from human
colon adenocarcinoma, Caco-2 cells express transporter
proteins, efflux proteins [apically localized P-glycoprotein (P-gp) efflux pump],8–10 and Phase II conjuga-
tion enzymes responsible for metabolic transformation
of test substances. Consequently, Caco-2 cells are the
most frequently used in vitro tool for prediction of drug
permeability and absorption potential,11,12 as they have
shown to be predictive of transcellular-based absorption
including passive diffusion, active transport, efflux and
paracellular transport.13,14 As an alternative to Caco-2
cells, MDCK cells with over expressed human multidrug-resistant gene (hMDR), MDCKII-hMDR1, have
become the method of choice to study bidirectional
transport of compounds and to evaluate the influence of
P-gp.15–17 This is due to the significantly shorter culture
time (3–5 vs. 21 days) which reduces cost and increases
throughput, low expression levels of transporter proteins
and low metabolic activity. In spite of their moderate
bioavailabilities in humans,18,19 some macrolide antibiotics20,21 are identified to be substrates/inhibitors for
the main drug efflux pump P-gp,22–26 which is known to
restrict intestinal permeation. Studies in rats and humans
have shown that P-gp inhibition with macrolides in
addition to the cytochrome P450 3A4 (CYP3A4) inhibition,27 can cause severe drug-drug interactions affecting oral bioavailability of co-administered drugs.28–30
* Author to whom correspondence should be addressed. (E-mail: [email protected])
324
D. Nožinić et al., Macrolide Transport Using in vitro Assays
O
O
OH
HO
O
O
O
O
OH
O
O
OH
Clarithromycin
HO
OH
HO
N
HO
O
O
O
O
O
O
O
O
OH
Roxithromycin
N
N
O
O
O
N
O
O
O
O
HO
O
O
OH
Azithromycin
nologies (Invitrogen, Carlsbad, CA, USA). Caco-2 cells
were obtained at passage 47 (European Collection of
Cell Cultures, Health Protection Agency Culture Collections, Salisbury, UK). MDCKII-hMDR1 cells were
obtained at passage 30 (Netherlands Cancer Institute,
Amsterdam, Netherlands).
Parallel Artificial Membrane Permeability Assays
(PAMPA)
N
N
N
N
HO
O
O
O
O
OH
O
O
Erythromycin
O
OH
N
HO
O
O
O
HO
OH
N
HO
O
O
O
O
HO
OH
O
N
O
O
O
Telithromycin
Figure 1. Macrolides used in this study.
Recent papers indicate that uptake transporter function
of organic anion transporting polypeptide (OATP) family should be considered as an additional potential mechanism of drug-drug interaction with macrolides.31,32
In the present study, five macrolide antibiotics
(erythromycin, clarithromycin, roxithromycin, azithromycin, and telithromycin) were characterized and compared according to their permeability through biological
membranes (Figure 1). Their permeabilities were studied using three different in vitro methods: PAMPA and
cellular based Caco-2 and MDCKII-hMDR1 assays,
with and without known P-gp inhibitors, in order to
evaluate their applicability to macrolides.
EXPERIMENTAL
Materials
Clarithromycin, azithromycin (Pliva, Zagreb, Croatia),
telithromycin (Sanofi-aventis, Paris, France), and testosterone (Steraloids, Wilton, NH, USA) were used in this
study. Erythromycin, roxithromycin, verapamil, propanolol hydrochloride, warfarin, dimethyl sulfoxide
(DMSO), phosphate-buffered saline (PBS), Hank's
balanced salt solution (HBSS), 4-(2-hydroxyethyl)-1piperazineethanesulfonic acid (HEPES), lucifer yellow,
acetonitrile (all from Sigma-Aldrich, Munich, Germany), lecithin [Lipoid E100 (Lipoid, Ludwigshafen,
Germany)], and n-dodecane (Merck, Darmstadt, Germany) were purchased. Amprenavir and 9,10-dihydro5-methoxy-9-oxo-N-[4-[2-(1,2,3,4-tetrahydro-6,7-dimethoxy2-isoquinolinyl)ethyl]phenyl]-4-acridine-carboxamide
(GF120918A) were acquired from our compound repository (GlaxoSmithKline, Stevenage, UK). Dulbecco's
Modified Eagles medium (DMEM), fetal bovine serum
(FBS), glutamax, nonessential amino acids (NEAA),
penicillin-streptomycin, Fungizone (amphotericin B),
trypsin-EDTA, and Dulbecco's Phosphate Buffered
Saline (DPBS) were purchased from Gibco Life TechCroat. Chem. Acta 83 (2010) 323.
DMSO stock solution of tested substance was diluted
with PBS buffer (10 mmol dm–3, pH = 7.4) to a final
concentration of 500 µmol dm–3 with 5 % (volume fraction, φ) of DMSO. The artificial membrane was prepared using a 1 % (φ) solution of lecithin in n-dodecane
and sonicated using a probe to ensure complete dissolution. Avoiding pipette tip contact with the membrane,
5 µl of the lecithin/n-dodecane mixture was pipetted
into each donor plate well (MultiScreen-IP PAMPA
assay plate, Millipore, Billerica, MA, USA). Immediately after application of the artificial membrane, 150 µl of
macrolide-containing donor solution was added to each
well of the donor plate. A volume of 300 µl of buffer
[5 % (φ) DMSO in 10 mmol dm–3 PBS, pH = 7.4] was
added to each acceptor plate well (PTFE Acceptor plate,
Millipore, Billerica, MA, USA). The macrolide-filled
donor plate was then placed into the acceptor plate to
create a “sandwich”, making sure the bottom of the
membrane is in contact with the buffer. Standard controls, used in the assays were warfarin (500 µmol dm–3)
and testosterone (100 µmol dm–3) as markers of low and
high permeability, respectively. The plate was incubated
at room temperature (≈ 23 °C) for 3 and 24 h with a
medium agitation. After incubation, macrolide concentrations in donor and acceptor plates were determined
by LC-MS/MS method. The permeability studies were
performed in duplicate for all tested substances, macrolides and standard controls, and the apparent permeability Papp in [nm s–1] was calculated16 according to equation
 substanceA,End 
  107
Papp  C  ln 1 
(1)

substanceeq 

where
C
VD  VA
VD  VA   area  t
(2)
and VD and VA are donor (D) and acceptor (A) solution
volumes [cm3], respectively; area is active surface area
of the membrane (i.e., area = 0.24 cm2 for the
membrane in assay plate); t is incubation time [s];
[substance]A,End is concentration [µmol dm–3] of tested
substance in the acceptor compartment at the completion of the assay; and [substance]eq is concentration
of tested substance at theoretical equilibrium (i.e.,
325
D. Nožinić et al., Macrolide Transport Using in vitro Assays
[substance]eq = 166 µM for the substances with initial
concentration of 500 µmol dm–3).33
Caco-2 Assay
Caco-2 cells were cultured34 under controlled atmosphere (37 °C, 95 % air, 5 % CO2, 90 % relative humidity), seeded in plastic flasks and maintained with complete medium [DMEM with glucose (4500 mg mL–1),
10 % (φ) FBS, 1 % (φ) glutamax, 1 % (φ) NEAA, 1 %
(φ) streptoycin (100 μg mL–1), and 1 % (φ) Fungizone
(amphotericin B)]. Culture medium was changed every
2 days and cultures were split (1:3 to 1:6) every 5 days
when they reached 85–90 % confluence, using 0.05 %
(φ) trypsin-EDTA. For transport studies, following
equilibration with the culture medium for 2 h in CO2
incubator, Caco-2 cells at passages 54–57 were seeded
onto collagen coated 6-well polycarbonate Costar
Transwell filter inserts (Corning, New York, NY, USA)
in culture medium at the density of 2×105 cells per mL.
The medium was changed the day after the seeding and
every other day (or three times per week) and Caco-2
cells were cultured on the filters for 20 to 22 days. On
the day of the transport experiment, the culture medium
was changed and 2 h thereafter the integrity of the cell
monolayers was estimated microscopically and by measuring transepithelial electrical resistance (TEER) by
epithelial voltmeter Millicell-ERS (Millipore, Billerica,
MA, USA), before and after the experiment. After removing the culture medium, the monolayers were incubated in the transport medium, for 2 h.
Propanolol hydrochloride, a marker of high permeability and transcellular transport, was used as a
control in the assay, dissolved in DMSO (10 mg mL–1)
and diluted to a final concentration of 100 μmol dm–3
using HBSS containing 12 mmol dm–3 HEPES adjusted
at pH = 7.4 for preparation of transport medium. All
tested substances were dissolved in DMSO (30 mmol
dm–3 stock solutions) and later diluted with transport
medium to a final concentration of 50 μmol dm–3. Each
substance was tested in triplicate in both directions
[apical-to-basolateral (AB) and basolateral-to-apical
(BA)] and in the presence and absence of verapamil, a
P-gp inhibitor (50 μmol dm–3). The amount of DMSO
applied to the cells did not exceed 1 % (φ). For AB
transport studies, the medium was removed from monolayers and the inserts were transferred to the new 6-well
plate. Each well contained prewarmed transport medium
(2.6 mL, 37 °C) in the lower compartment. Prewarmed
solutions of tested substances (1.5 mL, 37 °C) were added to the upper compartments. For BA transport studies,
prewarmed tested substances were added into the lower
compartments (2.6 mL, 37 °C) and prewarmed transport
medium was filled into the upper compartments (1.5 mL,
37 °C). In AB and BA bidirectional studies, samples
(100 μL) were taken at 0, 45, 90 and 135 min from the
acceptor compartments and immediately replenished
with the same volume of the transport medium. From
the donor compartments, aliquots (20 μL) were taken at
the beginning of the experiment (zero point) and at the
end of the experiment (135 min). During transport experiment, the plates were incubated at 37 °C and shaken at
55 rpm. In addition, analogue experiments with pH
gradient with pH = 6.5 and 7.4 in apical and basolateral
compartments, respectively, were done for all tested
substances. The concentrations of substances in the
collected samples from donor and acceptor wells were
determined by LC-MS/MS. The apparent permeability
Papp in [nm s–1] was calculated according to the equation
Papp 
dQ
 c0 1  area 1
dt
(3)
where dQ/dt is the linear appearance rate obtained from
the profile of the transported amount of the substrate
against the time [μg s–1]; c0 is a measured initial concentration in donor compartment [μg mL–1]; and area is the
membrane surface area of cell monolayer (i.e., area =
4.71 cm2 for 6-well plate). The ratio between Papp values
in both transport studies (AB and BA) was calculated
according to the expression
Papp ratio  BA/AB  
Papp (BA)
Papp (AB)
(4)
Mass balance MB was calculated by equation
MB / 
mD,End  mA,End
mD,0
 100
(5)
where m represents mass of tested substance in donor
(D) and acceptor (A) compartments at t = 0 min and at
the end of transport assay (End) and all values above
80 % were acceptable.
MDCKII-hMDR1 Assay
MDCKII-hMDR1 cells were grown in controlled
atmosphere (37 °C, 95 % air, 5 % CO2, 90 % relative
humidity) in DMEM with 10 % (φ) FBS, 1 % (φ) of
glutamax I, 1 % (φ) of penicillin (100 U mL–1) - streptomycin (100 μg mL–1) and 1 % (φ) of Fungizone (amphotericin B) in cell culture flasks. Culture medium was
changed every 2 days and cultures were split (1:3 to 1:6)
every 5 days when they reached 85–90 % confluence,
using 0.05 % (φ) trypsin-EDTA. Cells were prepared for
transport studies by pre-incubation of 24-well cell culture inserts (Millipore, Billerica, MA, USA) with culture medium (1 h, 37 °C) and then seeded in a concen-
Croat. Chem. Acta 83 (2010) 323.
326
tration of 0.3×106 cells per mL. The cells were fed with
fresh medium 24 h post seeding and cultured to confluence for 3 days before use. On the day of experiment, 2
h after feeding with the appropriate medium, the cell
monolayers were washed and equilibrated with transport
medium (DPBS, pH = 7.4) for 30 min (37 °C, 95 %
humidity). Test substance solution consisted of test
substance (5 μmol dm–3) in transport medium containing
lucifer yellow (100 μmol dm–3) and 1 % (φ) DMSO.
Transport assays were conducted using 400 and 800 μL
of apical and basolateral solution, respectively, i.e., for
AB study 400 μL of test substance solution was added
to upper compartment and 800 μL of transport medium
in the lower compartment. Monolayers were incubated
with the dose solution for 90 min at 37 °C under gentle
agitation. Apical and basolateral compartments were sampled at the end of the transport experiment and the tested
substance concentrations in both compartments were
determined by LC-MS/MS. Each substance was tested
three times in duplicate in both directions (AB and BA)
and in the presence and absence of GF120918A (2 μmol
dm–3), a specific P-gp inhibitor. Amprenavir (0.5 μmol
dm–3) was used as a positive control in the assay. Like
in Caco-2 assay, Papp in both directions (AB and BA),
and Pappratio for all tested substances and amprenavir
with and without GF120918A were determined using
the same equations [3] and [4]. Mass balance was calculated using expression [5] and all values in range of
80–120 % were acceptable. Lucifer yellow, a fluorescent marker for the paracellular pathway, was used as an
internal control in every monolayer to verify tight junction integrity during the assay. Cell monolayers with
lucifer yellow Papp > 30 nm s–1 were discarded.
LC-MS/MS Analysis
The LC-MS/MS system consisted of an Agilent 1100
LC Binary Pump (Agilent Technologies, Waldbronn,
Germany) with CTC Analytics HTS Pal autosampler
(CTC Analytics, Zwingen, Switzerland) and a SCIEX
API2000 or API3000 triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA, USA).
Chromatographic separation was performed using a
Phenomenex C18 guard column and Phenomenex
C18(2) (50×2.0 mm, 3 µm) column (Phenomenex, Torrance, CA, USA) for gradient method with acetonitrile/water (both acidified with 0.1 % (φ) formic acid) as
mobile phase, with flow rate of 0.3 mL min–1 at room
temperature, and with injection volume of 5 µl. Tested
substances were analysed using positive-ion mode with
the following multiple reaction monitoring (MRM)
transitions: erythromycin 735.5 (M+) → 577.5 (M+);
clarithromycin 748.5 (M+) → 590.3 (M+); roxithromycin
837.8 (M+) → 678.5 (M+); azithromycin 749.5 (M+) →
591.5 (M+); telithromycin 407.5 (M++) → 656.1 (M+);
warfarin 309.1(M+) → 163.3 (M+); propranolol 260.2
Croat. Chem. Acta 83 (2010) 323.
D. Nožinić et al., Macrolide Transport Using in vitro Assays
(M+) → 116.2 (M+); testosterone 289.3 (M+) → 97.2
(M+) and amprenavir 506 (M+) → 245.3 (M+). Samples
from PAMPA, Caco-2 and MDCKII-hMDR1 studies
were prepared by deproteinization with 3 volumes of
acetonitrile:methanol (φ = 0.67) containing internal
standard (clarithromycin or roxithromycin). Thereafter, samples were centrifuged (4000 rpm, 15 min,
4 °C) and an aliquot of supernatant was injected for
analysis.
Each tested substance sequence consisted of one
calibration curve and samples. Calibration curves were
designed to cover possible ranges of analyte concentrations in matrix: a) 11 dilutions in range of 0.195–200
μmol dm–3 for all tested substances in PAMPA; b) 10
dilutions in ranges of 0.1–50 μmol dm–3 and 0.2–100
μmol dm–3 for all tested macrolides and propranolol,
respectively, in Caco-2 assay; and c) 12 dilutions in ranges
of 0.0048–10 μmol dm–3 and 0.48 nmol dm–3–1 μmol dm–3
for all tested macrolides and amprenavir, respectively,
in MDCKII-hMDR1 assay. Linear regression analysis
with 1/x or 1/x2 weighing factor was chosen as the most
appropriate to determine the substance calibration
curves. The correlation coefficient in each calibration
curve was higher than 0.95. All back calculated values
used within each selected calibration curve were below
20 % of deviation. Peak areas of analyte and calibration
curves were determined using the Analyst software
version 1.4 (Applied Biosystems, Foster City, CA,
USA). Other calculations were performed in MS Excel
spreadsheet.
RESULTS
Papp values in PAMPA, obtained at pH = 7.4 with incubation time of 3 h at room temperature (≈ 23 °C), are
5.7 ±0.4
7.3 ± 0.7
warfarin
58.5 ±15.9
testosterone
10.2± 0.6
1.4 ±0.1
1.3 ±0.1
erythromycin
109.9±13.1
clarithromycin
37.8±1.7
65.5±21.8
roxithromycin
35.6± 1.5
13.1±0.5
10.1± 0.9
azithromycin
n.d.
0.13 ±0.03
telithromycin
0
30
3h
24 h
60 90 120 150 180
Papp / [nm s-1]
Figure 2. The apparent permeability coefficients (Papp/nm s–1)
in PAMPA at pH = 7.4 obtained by incubating tested substances
either for 3 or 24 h at room temperature (≈ 23 °C). Data is presented as mean value ± standard deviation from two replicates.
327
D. Nožinić et al., Macrolide Transport Using in vitro Assays
Table 1. The apparent permeabilities (Papp/nm s–1)* in Caco-2 cell assay from apical-to-basolateral (AB) and basolateral-to-apical
(BA) side on pH = 7.4 without and with inhibition of P-glycoprotein (P-gp) efflux pump by verapamil. The ratios between Papp in
BA and AB directions, [Pappratio(BA/AB)]
Substance
Papp (AB)
Noninhibited P-gp
Papp (BA)
Pappratio(BA/AB)
nm s –1
nm s –1
0.9  0.2
63.8  3.3
clarithromycin
38.4  4
roxithromycin
5.2  1.2
azithromycin
2.2  1.3
166.4  13.1
telithromycin
19.4  2.2
94.9  2.3
propanolol
121.8  4.5
94.2  3.1
0.8
erythromycin
Papp (AB)
Inhibited P-gp
Papp (BA)
Pappratio(BA/AB)
nm s –1
nm s –1
74.5
4.4  0.2
11.6  1.4
2.6
183.3  6.5
4.8
72.9  1.8
73.4  5.8
1.0
110.9  9.1
21.3
86.9  5.4
188.3  5.6
2.2
75
16.8  4.3
16.5  3.3
1.0
4.9
26.8  2.7
75.6  0.4
2.8
127.2  4.1
80.6  1.4
0.6
* data is presented as mean value ± standard deviation from three replicates.
displayed in Figure 2. The apparent permeability for
telithromycin could not be determined as the levels were
below the lower limit of quantification (0.195 µmol dm–3).
The highest PAMPA Papp values were obtained for clarithromycin, indicating high permeability, followed by
roxithromycin (moderate permeability), and low permeable azithromycin and erythromycin.
The influence of time on passive diffusion was examined by prolonging the incubation time from 3 to 24
h (Figure 2).16 For low permeable macrolides erythromycin, azithromycin and telithromycin, no substantial
differences in permeability with time were observed.
Moreover, an increase in incubation time did not contribute to an increase in telithromycin levels in the acceptor compartment significantly. For moderately permeable roxithromycin and highly permeable clarithromycin an increase in incubation time resulted in 1.8and 2.9-fold decreases in PAMPA Papp, respectively,
which classifies clarithromycin from highly to moderately permeable. Although prolonging the incubation
time produces different results, no significant change
was observed in the ranking of tested macrolides.
Caco-2 Papp values were determined for both di-
rections, AB and BA, and together with the respective
efflux ratio [Pappratio(BA/AB)] are listed in Table 1.
Erythromycin26 and azithromycin showed very low AB
permeability with Papp(AB) of 0.9 and 2.2 nm s–1, respectively, in comparison to the more permeable macrolides, roxithromycin (5.2 nm s–1), telithromycin (19.4
nm s–1) and clarithromycin (38.4 nm s–1). The addition
of verapamil (Table 1), resulted in increased Papp values
in the AB direction and in reduced Papp values in the BA
direction, respectively in all cases with the exception of
roxithromycin [Papp(BA) is 110.9 and 188.3 nm s–1
without and with P-gp inhibition, respectively]. Caco-2
efflux ratios with inhibited P-gp are significantly lower,
from 1 to 2.8, than without P-gp inhibition. The TEER
values varied slightly among different sets of plates and
only monolayers with TEER ≥ 300 Ω were used in
transport studies.
The effect of pH gradient, more precisely the effect of apical pH, on Caco-2 permeability was determined using pH = 6.5 and 7.4 solutions in apical compartment and pH = 7.4 solution in the basolateral compartment. Papp in both directions, AB and BA, and their
efflux ratio, are displayed in Figures 3a–c. While ery-
Figure 3. Effect of apical pH on the permeability of macrolides in Caco-2 cell assay. The apparent permeabilities (Papp/nm s–1) in a)
apical-to-basolateral (AB) direction, Papp(AB), b) in basolateral-to-apical (BA) direction, Papp(BA), and c) the ratio between Papp in BA
and AB directions [Pappratio(BA/AB)] obtained with (pH = 6.5 and 7.4 in apical and basolateral compartments, respectively) and without
pH gradient (pH = 7.4 in both compartments). Data is presented as mean value ± standard deviation from three replicates.
Croat. Chem. Acta 83 (2010) 323.
328
D. Nožinić et al., Macrolide Transport Using in vitro Assays
Table 2. The apparent permeabilities (Papp/nm s–1)* in MDCKII-hMDR1 cell assay from apical-to-basolateral (AB) and basolateral-to-apical (BA) side on pH = 7.4 without and with inhibition of P-glycoprotein (P-gp) efflux pump by GF120918A. The
ratios between Papp in BA and AB directions, [Pappratio(BA/AB)]
Substance
Papp (AB)
Noninhibited P-gp
Papp (BA)
Pappratio(BA/AB)
Papp (AB)
nm s –1
nm s –1
erythromycin
10.1 ± 0.3
128.7 ± 8.7
clarithromycin
3.3 ± 0.9
roxithromycin
3.7 ± 0.4
azithromycin
25.9 ± 4.3
109.7 ± 18.7
4.2
telithromycin
4.7 ± 0.8
358.5 ± 16.6
75.6
propanolol
43.7 ± 7.5
738.2 ± 9.9
16.9
284.8 ± 11.9
Inhibited P-gp
Papp (BA)
Pappratio(BA/AB)
nm s –1
nm s –1
12.7
9.7 ± 0.4
62.6 ± 16.4
6.4
548.6 ± 42.2
166
54.1 ± 3.2
159.1 ± 15.3
2.9
298 ± 7.5
80
6.2 ± 0.9
98.5 ± 14.5
16
35.9 ± 0.1
48.2 ± 9.9
1.3
59.3 ± 13.4
151.6 ± 42.1
2.6
537.5 ± 34.4
1.9
* data is presented as mean value ± standard deviation from three replicates.
thromycin and azithromycin exhibit decreased permeabilities at lower apical pH, the other three macrolides
showed an increased permeability (Figure 3c). The
changes in efflux ratios are primarily caused by a decrease in Papp(AB) at lower apical pH, observed for all
tested substances with the exception of azithromycin
(Figure 3a). On the contrary, at lower pH, a relatively
small decrease in Papp(BA) was observed for all but
roxithromycin (Figure 3b).
Papp values and the efflux ratios obtained in the
MDCKII-hMDR1 assay, with and without P-gp inhibition by GF120918A, are listed in Table 2. Although
Papp(AB) values for all five tested substances indicate
low to moderate permeability, as observed in the Caco-2
assay, the overall ranking differs. Clarithromycin, roxithromycin and telithromycin exhibit low transport in the
AB direction (3.3, 3.7, and 4.7 nm s–1, respectively),
whereas erythromycin (10.1 nm s–1) and azithromycin
(25.9 nm s–1) exhibit higher transport. Upon P-gp inhibition with GF120918A, Papp(AB) was increased for all
macrolides except erythromycin where the change was
negligible. Substantial increases in AB permeability
were observed for clarithromycin and telithromycin with
Papp(AB) of 54.1 and 59.3 nm s–1, respectively. Permeability in BA direction was reduced by P-gp inhibition
from 2 to 3.5 times for all of them. MDCKII-hMDR1
efflux ratios were significantly lower upon P-gp inhibition for all tested substances.
DISCUSSION
Transport studies using PAMPA, Caco-2 and MDCKIIhMDR1 assays were performed for a set of five commercially available macrolide drugs: 14-membered ring
macrolides erythromycin, clarithromycin, and roxithromycin, 15-membered ring macrolide azithromycin, and
for a novel class representative, ketolide, 14-membered
ring macrolide telithromycin.
PAMPA Papp describes the affinity of compounds
Croat. Chem. Acta 83 (2010) 323.
for transcellular passive diffusion in a system lacking
active and paracellular transport. The macrolides tested
are all weak bases with a protonated amine on the desosamine moiety at pH = 7.4. In addition, azithromycin
has the protonated nitrogen in the macrolide ring at the
same pH. Consequently, as being protonated at pH = 7.4,
the tested substances showed relatively low PAMPA
permeability. Our results, obtained after 3 h of incubation, classify erythromycin and azithromycin as low
permeable substances with the PAMPA Papp of 1.4 and
13.1 nm s–1, respectively (Figure 2). Moderately permeable roxithromycin and clarithromycin showed significantly higher PAMPA Papp (65.5 and 109.9 nm s–1,
respectively). These results are in line with their higher
oral bioavailability in humans (72–85 % and 55 %,
respectively)19 in comparison to erythromycin and azithromycin (20 % and 37 %, respectively).19,35 Although
they share similar physicochemical properties,36 a relatively small structural difference between erythromycin
and clarithromycin (hydroxy- vs. methoxy-substituent
on the C6 position, respectively) causes a significant
difference (79-fold increase) in their PAMPA permeability.
Although prolonging the incubation time from 3 to
24 h, did not significantly change the overall ranking of
tested substances, decreases in Papp values were observed for clarithromycin and roxithromycin. Decreased
Papp values could be explained by an increase of tested
substance concentration in the acceptor compartment
which comes closer to the theoretical equilibrium concentration ([substance]eq in Equation 1). Moreover, the
PAMPA assay classifies telithromycin as nonpermeable
as the concentrations of telithromycin were below the
lower limit of quantification after 3 h of incubation.
However, data from literature demonstrate that telithromycin has a good oral bioavailability in humans, very
similar to clarithromycin (57 vs. 55 %, respectively).19,21
In both cellular assays, Caco-2 and MDCKIIhMDR1, permeability measurements are made in the
329
D. Nožinić et al., Macrolide Transport Using in vitro Assays
apical-to-basolateral and basolateral-to-apical directions
with the apical side representing the gut lumen and the
basolateral side representing the systemic circulation.
The Caco-2 cell monolayer model mimics not only
transcellular passive diffusion, as PAMPA, but also
includes active transport, efflux and paracellular transport. However, due to macrolide size (Mr > 700), paracellular transport can be neglected in this study. The
influence of the pH gradient on the Caco-2 permeability
for tested macrolides is not straightforward (Figure 3).
Based on our results, the pH seems to affect not only the
passive diffusion but could also affect the affinity of
macrolides to behave as substrates for various efflux
pumps and/or transporters.
In the Caco-2 assay performed at pH = 7.4, all
tested macrolides showed lower permeability in the AB
direction in comparison to the BA direction (Table 1).
Moreover, erythromycin and azithromycin showed a
higher efflux ratios (74.5 and 75, respectively), when
compared to same values obtained for roxithromycin
(21.3), but especially for clarithromycin and telithromycin (4.8 and 4.9, respectively). Differences in AB permeabilities are likely due to different affinities for the
P-gp efflux system for the macrolides tested. It could be
postulated that transport of erythromycin and azithromycin is more influenced by active secretion and efflux
system than other three macrolides.
As previously reported in the literature, a comparison of efflux ratio generated in the presence and absence of a P-gp inhibitor (i.e. verapamil) is the method
of choice in identifying P-gp substrates.2 A compound is
typically considered to be a P-gp substrate when the
efflux ratio in the absence of inhibitor is greater than 2
and is at least halved in the presence of inhibitor.3
Moreover, inhibition of P-gp by verapamil generally
increased AB and decreased BA permeability.26 Accordingly, erythromycin, clarithromycin, roxithromycin,
and azithromycin could be classified as P-gp substrates
(Table 1). On the other hand, telithromycin does not
fulfil the second criteria, as efflux ratio decreased for 42 %
by P-gp inhibition. However, this result is in accordance
with previous findings in the literature that showed a
much lower extent of P-gp involvement in the overall
transport of telithromycin than for other macrolides.25
It has been proposed that a change in efflux ratio in the
presence of verapamil suggests that the greatest
decrease in efflux ratio is indicative of the lowest relative apparent affinity for P-gp.25 According to our
results obtained in the Caco-2 assay, the apparent P-gp
affinity is in following order: azithromycin < erythromycin < roxithromycin < clarithromycin < telithromycin.
One of the drawbacks of applied bidirectional
permeability Caco-2 assay is the fact that other transporters, uptake/efflux proteins and metabolic enzymes are
also functionally expressed in Caco-2 cells, therefore
ranking of analysed macrolides according to their P-gp
affinity could be ambiguous, at least. It raised the need
to perform comparable experiments using MDCKIIhMDR1 cell line with overexpression of human P-gp,
low expression levels of other transporter proteins and
low metabolic activity, in comparison to Caco-2 cells.
Our results show that MDCKII-hMDR1 permeability in AB direction is much lower than in the BA
direction for all tested macrolides (Table 2), which is in
line with data from the Caco-2 assay. However, the
observed permeabilities in the BA direction are generally higher in MDCKII-hMDR1 than in the Caco-2
assay. MDCKII-hMDR1 efflux ratios range from 4.2 to
166, due to different and in some cases very low AB
permeabilities. These results suggest a significant influence of the efflux mechanism on the permeability of
tested macrolides in MDCKII-hMDR1 assay, especially
in the cases of clarithromycin (166), roxithromycin (80),
and telithromycin (75.6). Differences in the permeability observed between Caco-2 and MDCKII-hMDR1
assays could be result of the difference in the relative
expression levels of total P-gp in each of them. Importantly, there is no data in literature that showed expression levels of P-gp or any transporters in MDCK cells
and this should be taken into account when interpreting
results of this assay.
The potency of P-gp inhibition in MDCKIIhMDR1 cells was assessed using verapamil and
GF120918A in the same experimental set-up (data not
shown). In these studies GF120918A demonstrated the
same inhibitory potency at a 25 times lower concentration and was therefore used for further studies. P-gp
inhibition with GF120918A in MDCKII-hMDR1 cells
reduced the amount of tested substance transported from
the cell to the apical side, thereby affecting the permeability in both directions: increasing Papp(AB) and decreasing Papp(BA) for all tested macrolides, but erythromycin. Interestingly, inhibition of P-gp with GF120918A
resulted in a decreased transport of erythromycin in the
BA direction, however, failed to influence its transport
in the AB direction. Similar effects have been reported
in the literature.25 Comparison of efflux ratio with and
without P-gp inhibition reveals a significant decrease
caused by P-gp inhibition, from 49 % for erythromycin
to 98 % for clarithromycin, classifying all these macrolides as P-gp substrates.
CONCLUSION
Our data demonstrate that the tested set of macrolides
behave as low to moderately permeable substances in all
three assays used. Results obtained in cellular assays,
Caco-2 and MDCKII-hMDR1, indicate involvement of
active transport by being P-gp substrates. In order to
Croat. Chem. Acta 83 (2010) 323.
330
D. Nožinić et al., Macrolide Transport Using in vitro Assays
assess the impact of P-pg efflux activity on the oral
absorption, in vivo studies in P-gp-deficient mice should
be considered. Based on our results we can conclude
that permeabilities obtained by PAMPA do not correlate
with values obtained in both cellular assays and that this
method is not appropriate for routine permeability
screening of macrolides, as it only takes into account
passive transport and likely inaccurately estimates the
intrinsic permeability.
Acknowledgements. The authors thank Vesna Munić, PhD, for
helpful and productive discussions during experimental work
described in this paper.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
S. Deferme, P. Annaert and P. Augustijns, Biotech. Pharm. Aspect 7 (2008) 182–215.
P. V. Balimane, Q.-H. Han and S. Chong, AAPS Journal 8
(2006) E1–E13.
E. H. Kerns, L. Di, S. Petusky, M. Farris, R. Ley and P. Jupp, J.
Pharm. Sci. 93 (2004) 1440–1453.
A. Avdeef, S. Bendels, L. Di, B. Faller, M. Kansy, K. Sugano
and Y. Yamauchi, J. Pharm. Sci. 96 (2007) 2893–2909.
A. Avdeef, Expert Opin. Drug Metab. Toxicol. 1 (2005)
325–342.
M. Kansy, A. Avdeef and H. Fischer, Drug Discov. Today.
Technol. 1 (2004) 349–355.
A. Avdeef and O. Tsinman, Eur. J. Pharm. Sci. 28 (2006) 43–50.
D. Balayssac, N. Authier, A. Cayre and F. Coudore, Toxicol.
Lett. 156 (2005) 319–329.
S. Eberl, B. Renner, A. Neubert, M. Reisig, I. Bachmakov, J.
König, F. Dörje, T. E. Mürdter, A. Ackermann, H. Dormann, K.
G. Gassmann, E. G. Hahn, S. Zierhut, K. Brune and M. F.
Fromm, Clin. Pharmacokinet. 46 (2007) 1039–1049.
J. P. Keogh and J. R. Kunta, Eur. J. Pharm. Sci. 27 (2006)
543–554.
R. B. Van Breemen and Y. Li, Expert Opin. Drug Metab.
Toxicol. 1 (2005) 175–185.
B. Press and D. Di Grandi, Curr. Drug Metab. 9 (2008)
893–900.
C. Masungi, J. Mensch, A. Van Dijck, C. Borremans, B. Willems, C. Mackie, M. Noppe and M. E. Brewster, Pharmazie 63
(2008) 194–199.
H. Sun, E. C. Y. Chow, S. Liu, Y. Du and K. S. Pang, Expert
Opin. Drug Metab. Toxicol. 4 (2008) 395–411.
T. T. Tran, A. Mittal, T. Gales, B. Maleeff, T. Aldinger, J. W.
Polli, A. Ayrton, H. Ellens and J. Bentz, J. Pharm. Sci. 93 (2004)
Croat. Chem. Acta 83 (2010) 323.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
2108–2123.
S. Carrara, V. Reali, P. Misiano, G. Dondio and C. Bigogno, Int.
J. Pharm. 345 (2007) 125–133.
P. Acharya, M. P. O'Connor, J. W. Polli, A. Ayrton, H. Ellens
and J. Bentz, Drug Metab. Dispos. 36 (2008) 452–460.
G. Foulds, R. M. Shepard and R. B. Johnson, J. Antimicrob.
Chemother. 25 (Suppl. A) (1990) 73–82.
R. Jain and L. H. Danziger, Curr. Pharm. Design 10 (2004)
3045–3053.
G. G. Zhanel, M. Dueck, D. J. Hoban, L. M. Vercaigne, J. M.
Embil, A. S . Gin and J. A. Karlowsky, Drugs 61 (2001) 443–98.
M. Zeitlinger, C. C. Wagner and B. Heinisch, Clin. Pharmacokinet. 48 (2009) 23–38.
C. Seral, J.-M. Michot, H. Chanteux, M.-P. Mingeot-Leclercq, P.
M. Tulkens and F. Van Bambeke, Antimicrob. Agents Chemother. 47 (2003) 1047–1051.
M. Sugie, E. Asakura, Y. L. Zhao, S. Torita, M. Nadai, K. Baba,
K. Kitaichi, K. Takagi, K. Takagi and T. Hasegawa, Antimicrob.
Agents Chemother. 48 (2004) 809–814.
S. Yamaguchi, Y. L. Zhao, M. Nadai, H. Yoshizumi, X. Cen, S.
Torita, K. Takagi, K. Takagi and T. Hasegawa, Antimicrob.
Agents Chemother. 50 (2006) 80–87.
J. I. Pachot, R. P. Botham and K. D. Haegele, J. Pharm. Pharmaceut. Sci. 6 (2003) 1–12.
H. Saito, Y. Fukasawa, Y. Otsubo, K. Yamada, H. Sezaki and S.
Yamashita, Pharm. Res. 17 (2000) 761–765.
J. F. Westphal, Br. J. Clin. Pharmacol. 50 (2000) 285–295.
K. Shakeri-Nejad and R. Stahlmann, Expert Opin. Pharmacother. 7 (2006) 639–651.
J. Rengelshausen, C. Goggelmann, J. Burhenne, K. D. Riedel, J.
Ludwig, J. Weiss, G. Mikus, I. Walter-Sack and W. E. Haefeli,
Br. J. Clin. Pharmacol. 56 (2003) 32–38.
S. Ito, R. Nasu, M. Tsujimoto, H. Murakami, H. Ohtani and Y.
Sawada, Biopharm. Drug Dispos. 28 (2007) 113–123.
E. Garver, E. D. Hugger, S. P. Shearn, A. Rao, P. A. Dawson,
C. B. Davis and C. Han, Drug Metab. Dispos. 36 (2008)
2492–2498.
A. Seithel, S. Eberl, K. Singer, D. Auge, G. Heinkele, N. B.
Wolf, F. Dörje, M. F. Fromm and J. König, Drug Metab. Dispos.
35 (2007) 779–786.
K. Sugano, H. Hamada, M. Machida and H. Ushio, J. Biomol.
Screen. (2001) 6189–6196.
J. G. Gao, E. D. Hugger, M. S. Beck-Westernmeyer and R. T.
Borchart, Curr. Protoc. Pharmacol. 7 (2000) 7.2.1–7.2.23
J. Lakritz and D. W. Wilson, AAEP Proceedings 43 (1997)
83–86.
J. W. McFarland, C. M. Berger, S. A. Froshauer, S. F. Hayashi,
S. J. Hecker, B. H. Jaynes, M. R. Jafson, B. J. Kamicker, C. A.
Lipinski, K. M. Lundy, C. P. Reese and C. B. Vu, J. Med. Chem.
40 (1997) 1340–1346.
331
D. Nožinić et al., Macrolide Transport Using in vitro Assays
SAŽETAK
Utvrđivanje transporta makrolida putem PAMPA, Caco-2
i MDCKII-hMDR1 testova
Danijela Nožinić,a Astrid Milić,a Lara Mikac,b Jovica Ralić,a Jasna Padovan,a
i Roberto Antolovićc
a
GlaxoSmithKline centar za istraživanje Zagreb, Ltd., Prilaz baruna Filipovića 29, 10000 Zagreb,
Croatia
b
Zavod za fiziku materijala, Institut Ruđer Bošković, Bijenička 54, 10000 Zagreb, Croatia
c
Odjel za biotehnologiju, Sveučilište u Rijeci, S. Krautzeka bb, 51000 Rijeka, Croatia
Transport komercijalnih makrolidnih antibiotika eritromicina, klaritromicina, roksitromicina, azitromicina i telitromicina proučavan je PAMPA testom pasivne difuzije te Caco-2 i MDCKII-hMDR1 testovima aktivnog transporta molekula kroz membranu. Proučavan je i utjecaj membranskog transportera P-glikoproteina (P-gp) na transport makrolida. U Caco-2 i MDCKII-hMDR1 testovima bez P-gp inhibicije opažena je smanjena permeabilnost
(Papp) iz apikalne prema bazolateralnoj (AB) nego li iz bazolateralne prema apikalnoj strani (BA) membrane
[omjer permeabilnosti, Pappratio(BA/AB) > 4,2]. Uslijed P-gp inhibicije Papp(AB) se povećala, dok se Papp(BA)
smanjila. Opažene Papp vrijednosti, omjeri permeabilnosti, kao i značajno smanjenje istih omjera uzrokovano P-gp
inhibicijom (> 50 %), svrstava ovdje testirane makrolide u P-gp supstrate. Rezultati dobiveni PAMPA testom
upućuju na zaključak da ovaj model nije dobar izbor za razvrstavanje makrolida prema njihovoj permeabilnosti,
kako radi nepostojanja P-gp tako i/ili radi nepostojanja drugih transportera potencijalno uključenih u transport
makrolida.
Croat. Chem. Acta 83 (2010) 323.

Download Report